a-mem-mcp: Build Evolving Memory for Coding Agents
Summary
A-MEM is an MCP server and Python library that stores agent knowledge as an evolving, connected memory graph. It suits coding-agent users who need to retrieve and build on project context across sessions.
At a glance
- Language
- Python
- License
- MIT
- Stars
- 35
- Forks
- 3
- Added to OSRepos
- August 17, 2026
- Last analyzed
- October 4, 2026
Topics
Click on any tag to explore related repositories
Use at your own risk
OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of code from these repositories is the user's own responsibility. Always review the repository, source code, dependencies, licenses, and security implications before running or installing anything. OSRepos is not responsible for issues, damages, or losses resulting from third-party repositories.
Overview
A-MEM gives coding agents a persistent memory system that connects related notes instead of treating each saved fact as an isolated vector-store entry. When new knowledge is added, it uses an LLM to extract context and metadata, find related memories, and evolve their links and descriptions.
The project is intended for agents that need to recall project-specific decisions, implementation details, or patterns across sessions. It provides MCP tools for agent use and a Python API for direct integration. The README reports testing with Claude Code, with support for other MCP-compatible agents planned.
Key Features
- Stores notes with LLM-generated keywords, tags, and context.
- Builds and updates relationships between semantically similar memories.
- Combines semantic search with graph traversal to explore connected knowledge.
- Supports lightweight metadata search followed by full-note reads to manage context use.
- Provides tools for adding, searching, reading, updating, and deleting notes, plus checking asynchronous task status.
- Offers a Python API as well as an MCP server interface.
- Uses ChromaDB for persistence and supports configurable LLM backends, including OpenAI, Ollama, SGLang, and OpenRouter.
Use Cases
- Claude Code users can preserve architecture notes and implementation decisions between coding sessions.
- Development teams can make recurring project conventions searchable when an agent works across a codebase.
- Agent developers can add persistent, connected memory through MCP rather than implementing note storage and retrieval themselves.
- Python application developers can use the memory system directly when they need agent memory outside an MCP client.
Project Facts
- Language: Python
- License: MIT
- Stars: 35
- Forks: 3
- Topics: none listed
- Archived: no
Getting Started
Install the package with:
pip install a-mem
See the README for Claude Code setup, configuration, and usage details.
Alternatives
- agentmemory: Agentmemory captures coding work through integrations and shares searchable memory via MCP and REST, rather than centering on an evolving linked memory graph.
- Agent-Memory: Agent-Memory focuses on automatic context capture and recall for MCP-compatible coding agents, rather than an evolving connected memory graph.
- Memary: Memary combines knowledge-graph memory with user-focused personalization for autonomous agents, rather than targeting coding-agent project context specifically.
Considerations
- The README identifies Claude Code as the currently tested MCP-compatible agent; broader agent support is planned, not established.
- Using the default OpenAI backend requires an API key. The README also documents alternative backends, including local Ollama.
- Memory is project-specific by default, stored in
./chroma_db. A different path is needed to share it across projects. - Adding memories invokes an LLM and is asynchronous through the MCP tool interface, so clients may need to check task completion.
- Repository metadata shows an early project, created in January 2026, with 35 stars and no listed open issues. The supplied README describes capabilities but does not provide independent performance or reliability results.
Source repository
Open the original repository on GitHub.
15 counted GitHub visits
Related repositories
Similar repositories that may be relevant next.
web-design: Create Consistent Web Pages with a Claude Code Skill
October 3, 2026
web-design is a Claude Code skill that turns product briefs, reference URLs, or screenshots into an editable design specification before generating web code. It is suited to developers and designers who want a repeatable, spec-led workflow for building consistent pages.
oomwoo: Build a DIY Robot Vacuum
October 2, 2026
OOMWOO is a planned, hackable robot vacuum built around Raspberry Pi, ROS2 and 2D LiDAR. It is aimed at makers who want to build and customize a locally controlled vacuum, but its hardware and build instructions are still in development.
shepherd: Supervise Agents with Reversible Execution Traces
October 2, 2026
Shepherd records agent work as inspectable, reversible execution traces and keeps changes as proposals for review. It is aimed at developers building systems that supervise, replay, or manage the work of other agents.
agent-anvil: Test AI Agent Tool Use in CI
October 1, 2026
Agent Anvil evaluates tool-using AI agents through scenario-based runs, trace checks, and optional semantic grading. It helps teams catch unsafe or incorrect tool behavior before it reaches production.